In the therapy of certain conditions it is preferable that a drug be delivered only to a particular cellular subpopulation. For example the use of drugs in the treatment of cancer is limited by the inability of the cytotoxic drug to differentiate between cells exhibiting normal cell division and those exhibiting neoplastic division. Hence the therapy is not targeted to a clinically acceptable extent and healthy cells are exposed to cytotoxin. Conjugation of a drug to an antibody, preferably a monoclonal antibody (mAb), allows the targeting of the drug to a particular cellular subpopulation expressing the antigenic determinant to which the targetting antibody binds. However factors, such as the inability of the conjugate to penetrate the relevant tissue, poor release of the drug from the antigen bound conjugate and the limitation placed on the amount of drug which can be delivered by the number of available antibody-binding sites, have limited the effectiveness of this approach.
Avoidance of such limitations led to the concept of targeting conjugates of antibodies and enzymes capable of converting relatively non-cytotoxic `prodrugs` into low molecular weight cytotoxins at the antibody-binding site. This general concept was disclosed by Rose in European Patent application 84302218.7. Bagshawe and collaborators have referred to the concept as ADEPT, Antibody Directed Enzyme Prodrug Therapy. (Bagshawe K. D. et.al., Br. J. Cancer [1987] 56, 531-532, Bagshawe K. D. el al., Br. J. Cancer [1988] 50, 700-703 and WO 90/10140). In this way one conjugate could generate a proportionately larger amount of drug at the target site by repeated rounds of enzymatic catalysis of prodrug activation.
EP 382 411 describes ADEPT wherein a prodrug may be converted to a cytotoxic agent by enzymes including beta-lactamases isolated from various micro-organisms, L-pyroglutamate aminopeptidase, beta-galactosidase, D-amino acid peptidase, isoenzymes of alkaline phosphatase and various carboxypeptidases.
WO 91/11201 describes ADEPT wherein a prodrug may be enzymatically cleaved to generate cyanide by .beta.-glycosidases or .beta.-glucosidases, generally of plant origin.
EP 302 473 describes ADEPT wherein the enzyme alkaline phosphatase may be used to cleave novel prodrugs of mitomycin, penicillin V amidase may be used to cleave novel prodrugs of adriamycin, or cytosine deaminase may be used with the prodrug 5-fluorocytosine.
EP 484 870 describes ADEPT wherein .beta.-lactamase may be used to activate a cephalosporin prodrug to yield a cytotoxic nitrogen mustard.
WO 88/07378 describes ADEPT wherein benzoic acid nitrogen mustard glutamides may be converted to the nitrogen mustard under the action of carboxypeptidases.
Vitols, K. S., el al, Pteridines 3, [1992], 125-126, discloses a number of MTX-amino acid prodrugs which may be activated by carboxypeptidase-mAb conjugates as part of ADEPT.
WO 91/09134 discloses bispecific hybrid mAbs for use in ADEPT wherein the mAb has specificities against both human cancer cell antigens and a prodrug-activating enzyme.
All previously disclosed methods of ADEPT may be divided into two categories: those which employ human enzymes and those which employ non-human (eg. bacterial) enzymes to activate the relevant prodrug. Both strategies retain inherent problems which limit their potential to provide effective therapy. Use of a human enzyme results in instability of the associated prodrug in vivo, as it may be activated at sites distant to the target site where endogenous human enzymatic activity may occur naturally. Clearly this will also have the highly undesirable effect of generating potentially cytotoxic compounds in non-targetted areas of the body with possibly lethal consequences. The use of a non-human enzyme permits the associated prodrug, which is only activated by the non-human enzymatic activity, to avoid activation by endogenous enzymes and so remain stable in vivo until converted to drug at the target site. However, such a non-human enzyme may elicit an inmnune response when introduced in vivo and antibodies generated to the enzyme will limit or destroy its ability to activate the prodrug.
WO 90/07929 identifies the desideratum of a non-endogenous catalytic activity being provided by a non-immunogenic enzyme but teaches only that this may be achieved by the use of "genetically conserved" enzymes or those from a "genetically similar species". However it is not taught how this may be accomplished.
It has now been found that the apparent conflict arising from the need to provide a prodrug which is stable in vivo and yet activated by a non-immunogenic enzyme may be resolved by generation of a mutant mammalian enzyme which retains catalytic activity but possesses a novel substrate specificity. The associated prodrug may be activated by the catalytic activity of the mutant enzyme but since the substrate specificity of the mutant enzyme is not a naturally occurring one, the prodrug remains stable in vivo until converted to drug at the target site. The ability to use a non-immunogenic enzyme according to the present invention provides the further advantage that repeated rounds of therapy may be administered. This is in contrast to known processes for ADEPT, in which the initial introduction of the enzyme to the system elicits an immune response which effectively precludes further treatment with the same enzyme as this will be removed from the body by an immune reaction `primed` during the first round of therapy.